Programming with Python Project 2

Marketing Data Analysis

Domain

Marketing, Retail

Problem Statement

• Prepare the given dataset for further analysis on the effectiveness of the marketing campaigns.
• Visualize the pertinent data fields.

Q1 Import the necessary libraries and load the dataset and display 5 random samples.Check the info of the data and write your findings.

# Importing the libraries required.
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import plotly.express as px

%matplotlib inline

# Setting display attributes for better readability of the dataframe.
import warnings
warnings.filterwarnings('ignore')
pd.options.display.float_format = '{:.2f}'.format
pd.set_option('precision', 2)

pd.set_option('display.max_rows', 500)
pd.set_option('display.max_columns', 500)
pd.set_option('display.width', 1000)

# Loading the dataset
df = pd.read_csv('marketing_data.csv')

# Displaying a random sample of 5 observations.
df.sample(5)

	ID	Year_Birth	Education	Marital_Status	Income	Kidhome	Teenhome	Dt_Customer	Recency	MntWines	MntFruits	MntMeatProducts	MntFishProducts	MntSweetProducts	MntGoldProds	NumDealsPurchases	NumWebPurchases	NumCatalogPurchases	NumStorePurchases	NumWebVisitsMonth	AcceptedCmp3	AcceptedCmp4	AcceptedCmp5	AcceptedCmp1	AcceptedCmp2	Response	Complain	Country
1141	8091	1956	Graduation	Married	$63,943.00	0	1	9/2/12	50	423	184	368	13	97	21	1	6	4	6	5	0	0	0	0	0	0	0	IND
1460	5966	1974	PhD	Married	$45,207.00	1	1	11/10/12	64	203	0	10	0	0	2	5	3	1	6	6	0	1	0	0	0	0	0	SP
1044	3933	1980	Graduation	Together	$44,010.00	1	0	10/13/12	46	186	36	234	86	72	48	6	10	2	6	9	0	0	0	0	0	0	0	SA
682	2631	1982	Graduation	Together	$71,853.00	0	0	5/8/13	29	358	108	413	141	97	32	1	2	8	6	1	0	0	0	0	0	0	0	SP
482	2666	1972	Master	Married	$76,234.00	0	1	2/6/14	21	519	50	167	130	0	41	1	8	3	11	3	0	1	0	0	0	0	0	CA

df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 2240 entries, 0 to 2239
Data columns (total 28 columns):
 #   Column               Non-Null Count  Dtype 
---  ------               --------------  ----- 
 0   ID                   2240 non-null   int64 
 1   Year_Birth           2240 non-null   int64 
 2   Education            2240 non-null   object
 3   Marital_Status       2240 non-null   object
 4    Income              2216 non-null   object
 5   Kidhome              2240 non-null   int64 
 6   Teenhome             2240 non-null   int64 
 7   Dt_Customer          2240 non-null   object
 8   Recency              2240 non-null   int64 
 9   MntWines             2240 non-null   int64 
 10  MntFruits            2240 non-null   int64 
 11  MntMeatProducts      2240 non-null   int64 
 12  MntFishProducts      2240 non-null   int64 
 13  MntSweetProducts     2240 non-null   int64 
 14  MntGoldProds         2240 non-null   int64 
 15  NumDealsPurchases    2240 non-null   int64 
 16  NumWebPurchases      2240 non-null   int64 
 17  NumCatalogPurchases  2240 non-null   int64 
 18  NumStorePurchases    2240 non-null   int64 
 19  NumWebVisitsMonth    2240 non-null   int64 
 20  AcceptedCmp3         2240 non-null   int64 
 21  AcceptedCmp4         2240 non-null   int64 
 22  AcceptedCmp5         2240 non-null   int64 
 23  AcceptedCmp1         2240 non-null   int64 
 24  AcceptedCmp2         2240 non-null   int64 
 25  Response             2240 non-null   int64 
 26  Complain             2240 non-null   int64 
 27  Country              2240 non-null   object
dtypes: int64(23), object(5)
memory usage: 490.1+ KB

df.shape

(2240, 28)

Inferences:

• The dataset has a total of 2240 rows and 28 columns.
• The datatypes for the columns are correctly assigned, except for income and dt_customer.
• The only null values are in the income field.
• Education , Marital Status and Country are the only categorical datatypes, the remaining are quantitative variables.

Q2 Check the following using an appropriate method and write your findings.

• Check how spread out or varied your data set is.
• Check where the middle 50% of your data lies.
• Check boundaries for the lower, middle and upper quarters of data.

# Checking the summary of the quantitative variables.
df.describe().T

	count	mean	std	min	25%	50%	75%	max
ID	2240.00	5592.16	3246.66	0.00	2828.25	5458.50	8427.75	11191.00
Year_Birth	2240.00	1968.81	11.98	1893.00	1959.00	1970.00	1977.00	1996.00
Kidhome	2240.00	0.44	0.54	0.00	0.00	0.00	1.00	2.00
Teenhome	2240.00	0.51	0.54	0.00	0.00	0.00	1.00	2.00
Recency	2240.00	49.11	28.96	0.00	24.00	49.00	74.00	99.00
MntWines	2240.00	303.94	336.60	0.00	23.75	173.50	504.25	1493.00
MntFruits	2240.00	26.30	39.77	0.00	1.00	8.00	33.00	199.00
MntMeatProducts	2240.00	166.95	225.72	0.00	16.00	67.00	232.00	1725.00
MntFishProducts	2240.00	37.53	54.63	0.00	3.00	12.00	50.00	259.00
MntSweetProducts	2240.00	27.06	41.28	0.00	1.00	8.00	33.00	263.00
MntGoldProds	2240.00	44.02	52.17	0.00	9.00	24.00	56.00	362.00
NumDealsPurchases	2240.00	2.33	1.93	0.00	1.00	2.00	3.00	15.00
NumWebPurchases	2240.00	4.08	2.78	0.00	2.00	4.00	6.00	27.00
NumCatalogPurchases	2240.00	2.66	2.92	0.00	0.00	2.00	4.00	28.00
NumStorePurchases	2240.00	5.79	3.25	0.00	3.00	5.00	8.00	13.00
NumWebVisitsMonth	2240.00	5.32	2.43	0.00	3.00	6.00	7.00	20.00
AcceptedCmp3	2240.00	0.07	0.26	0.00	0.00	0.00	0.00	1.00
AcceptedCmp4	2240.00	0.07	0.26	0.00	0.00	0.00	0.00	1.00
AcceptedCmp5	2240.00	0.07	0.26	0.00	0.00	0.00	0.00	1.00
AcceptedCmp1	2240.00	0.06	0.25	0.00	0.00	0.00	0.00	1.00
AcceptedCmp2	2240.00	0.01	0.11	0.00	0.00	0.00	0.00	1.00
Response	2240.00	0.15	0.36	0.00	0.00	0.00	0.00	1.00
Complain	2240.00	0.01	0.10	0.00	0.00	0.00	0.00	1.00

# Checking the total product sales
df.iloc[:, 10:16].sum()

MntFruits             58917
MntMeatProducts      373968
MntFishProducts       84057
MntSweetProducts      60621
MntGoldProds          98609
NumDealsPurchases      5208
dtype: int64

# Checking the number of purchases from different sales channels.
df.iloc[:,17:21].sum()

NumCatalogPurchases     5963
NumStorePurchases      12970
NumWebVisitsMonth      11909
AcceptedCmp3             163
dtype: int64

# Checking the summary of the categorical variables.
df.describe(include='O')

	Education	Marital_Status	Income	Dt_Customer	Country
count	2240	2240	2216	2240	2240
unique	5	8	1974	663	8
top	Graduation	Married	$7,500.00	8/31/12	SP
freq	1127	864	12	12	1095

df['Education'].value_counts().plot(kind='barh', figsize=(6,4), edgecolor=(0,0,0),color='blue', title='Education')

<AxesSubplot:title={'center':'Education'}>

df['Marital_Status'].value_counts().plot(kind='barh', figsize=(6,4), edgecolor=(0,0,0),color='green', title='Marital Status')

<AxesSubplot:title={'center':'Marital Status'}>

Inferences

• The median year of birth for the customers is 1970, the earlier year of birth is 1893 while the latest is 1996.
• The year of birth of 1893 is unlikely and we can imputate it later if impacts our analysis.
• The majority of the customers do not have kids or teens at home.
• Amongst the different products, Wines have the highest average number of sales, the lowest is for fruits.
• The majority of the customers are graduates and married.
• Since the income data field is not correctly formatted we will next analyze the income distribution.

Q3 Check for any missing values in the dataset and handle them using an appropriate method.

# Checking null values.
df.isna().sum()

ID                      0
Year_Birth              0
Education               0
Marital_Status          0
 Income                24
Kidhome                 0
Teenhome                0
Dt_Customer             0
Recency                 0
MntWines                0
MntFruits               0
MntMeatProducts         0
MntFishProducts         0
MntSweetProducts        0
MntGoldProds            0
NumDealsPurchases       0
NumWebPurchases         0
NumCatalogPurchases     0
NumStorePurchases       0
NumWebVisitsMonth       0
AcceptedCmp3            0
AcceptedCmp4            0
AcceptedCmp5            0
AcceptedCmp1            0
AcceptedCmp2            0
Response                0
Complain                0
Country                 0
dtype: int64

# Formatting the Income column for the correct data type.
df.rename(columns = {' Income ':'Income'}, inplace = True)
df['Income'] = df['Income'].str.replace(',','')
df['Income'] = df['Income'].str.replace('$','').astype(float)

# Checking the income distribution
sns.displot(df['Income'], kde=True)

<seaborn.axisgrid.FacetGrid at 0x2553e1479d0>

plt.figure(figsize=(10,5))
sns.distplot(df.Income, kde=False)
plt.ylabel('Count')
plt.title('Distribution of Income', size=15)

Text(0.5, 1.0, 'Distribution of Income')

# Checking the numerical summary
df['Income'].describe()

count     2216.00
mean     52247.25
std      25173.08
min       1730.00
25%      35303.00
50%      51381.50
75%      68522.00
max     666666.00
Name: Income, dtype: float64

# Imputing missing values
df['Income'].fillna((df['Income'].median()), inplace=True)

Missing value imputation

• Since the income data is right skewed , we used median value for imputation.

Q4 Check for any presence of special characters in any variables. If present, clean/replace and change the datatype of the variable if required.

# As described earlier converting the DT_Customer column to the pandas datetime format.
df["Dt_Customer"]= pd.to_datetime(df["Dt_Customer"])
df["Dt_Customer"]

0      2014-06-16
1      2014-06-15
2      2014-05-13
3      2014-05-11
4      2014-04-08
          ...    
2235   2013-03-07
2236   2013-01-22
2237   2012-12-03
2238   2012-11-29
2239   2012-09-01
Name: Dt_Customer, Length: 2240, dtype: datetime64[ns]

df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 2240 entries, 0 to 2239
Data columns (total 28 columns):
 #   Column               Non-Null Count  Dtype         
---  ------               --------------  -----         
 0   ID                   2240 non-null   int64         
 1   Year_Birth           2240 non-null   int64         
 2   Education            2240 non-null   object        
 3   Marital_Status       2240 non-null   object        
 4   Income               2240 non-null   float64       
 5   Kidhome              2240 non-null   int64         
 6   Teenhome             2240 non-null   int64         
 7   Dt_Customer          2240 non-null   datetime64[ns]
 8   Recency              2240 non-null   int64         
 9   MntWines             2240 non-null   int64         
 10  MntFruits            2240 non-null   int64         
 11  MntMeatProducts      2240 non-null   int64         
 12  MntFishProducts      2240 non-null   int64         
 13  MntSweetProducts     2240 non-null   int64         
 14  MntGoldProds         2240 non-null   int64         
 15  NumDealsPurchases    2240 non-null   int64         
 16  NumWebPurchases      2240 non-null   int64         
 17  NumCatalogPurchases  2240 non-null   int64         
 18  NumStorePurchases    2240 non-null   int64         
 19  NumWebVisitsMonth    2240 non-null   int64         
 20  AcceptedCmp3         2240 non-null   int64         
 21  AcceptedCmp4         2240 non-null   int64         
 22  AcceptedCmp5         2240 non-null   int64         
 23  AcceptedCmp1         2240 non-null   int64         
 24  AcceptedCmp2         2240 non-null   int64         
 25  Response             2240 non-null   int64         
 26  Complain             2240 non-null   int64         
 27  Country              2240 non-null   object        
dtypes: datetime64[ns](1), float64(1), int64(23), object(3)
memory usage: 490.1+ KB

Conclusion

• The only datatype which was not formatted correctly was DT_Customer and we will assign it to the date time format.
• Our dataset is now cleaned and does not have any missing or invalid values.

Q5 The Marketing Manager wants to know the 'Age' of the customers. Extract the feature age from the given dataset and display the statistical summary of the age.

# Inserting a new column for age and obtaining the value by subtracting the latest completed year i.e 2021 from the year of birth.
df.insert(2, 'age', (2021 - df['Year_Birth']))
df.head()

	ID	Year_Birth	age	Education	Marital_Status	Income	Kidhome	Teenhome	Dt_Customer	Recency	MntWines	MntFruits	MntMeatProducts	MntFishProducts	MntSweetProducts	MntGoldProds	NumDealsPurchases	NumWebPurchases	NumCatalogPurchases	NumStorePurchases	NumWebVisitsMonth	AcceptedCmp3	AcceptedCmp4	AcceptedCmp5	AcceptedCmp1	AcceptedCmp2	Response	Complain	Country
0	1826	1970	51	Graduation	Divorced	84835.00	0	0	2014-06-16	0	189	104	379	111	189	218	1	4	4	6	1	0	0	0	0	0	1	0	SP
1	1	1961	60	Graduation	Single	57091.00	0	0	2014-06-15	0	464	5	64	7	0	37	1	7	3	7	5	0	0	0	0	1	1	0	CA
2	10476	1958	63	Graduation	Married	67267.00	0	1	2014-05-13	0	134	11	59	15	2	30	1	3	2	5	2	0	0	0	0	0	0	0	US
3	1386	1967	54	Graduation	Together	32474.00	1	1	2014-05-11	0	10	0	1	0	0	0	1	1	0	2	7	0	0	0	0	0	0	0	AUS
4	5371	1989	32	Graduation	Single	21474.00	1	0	2014-04-08	0	6	16	24	11	0	34	2	3	1	2	7	1	0	0	0	0	1	0	SP

df['age'].describe()

count   2240.00
mean      52.19
std       11.98
min       25.00
25%       44.00
50%       51.00
75%       62.00
max      128.00
Name: age, dtype: float64

Conclusion

• The median age of our customer is 51 and since the average age is similar we can assume the ages are normally distributed.

Q6 The Marketing manager wants to understand the total amount spent on various products so that we can find what percentage of the amount is spent on which product.

• Find out the total amount spent by a customer.
• Display the Percentage of the amount spent on Wines and other products.

# Creating a new column for the total purchase amount.
df.insert(16, 'total', 0)

# Adding the total for different products to obtain the total purchase amount.
df['total'] = df.iloc[:, 10:17].sum(axis = 1)

# Creating a subsection of the dataframe to calculate the proportion of different products to the total purchase amount.
products = df.iloc[:, 10:17]

products = products.sum().to_frame()
products = (products / products.loc['total']) * 100
products.columns = ['Percentage']
print(products)

                  Percentage
MntWines               50.17
MntFruits               4.34
MntMeatProducts        27.56
MntFishProducts         6.19
MntSweetProducts        4.47
MntGoldProds            7.27
total                 100.00

# Creating a pie chart to show the proportion of purchase amount for the different products.
products = products.drop(['total'])
colors = sns.color_palette('pastel')[0:6]
labels = ['Wines', 'Fruits', 'Meat Products', ' Fish Products', 'Sweet Products', 'Gold']
plt.pie(products['Percentage'], labels = labels, colors = colors, autopct='%.0f%%')
plt.show()

Q7 Being an Analyst understands the total number of purchases made through different channels which can help find the percentage these channels contribute.

• Find out the total purchases done by a customer through different channels.
• Display the percentage of the store and other channels’ contribution to the total purchases.

# Creating a new column for the total number of purchases.
df.insert(21, 'total_purchase', 0)

# Calculating the total number of purchases.
df['total_purchase'] = df.iloc[:, 17:21].sum(axis = 1)

# Creating a subsection of the dataframe to calculate the percentage for the different sales channels.
sales_channel = df.iloc[:, 17:22]

sales_channel = sales_channel.sum().to_frame()
sales_channel = (sales_channel / sales_channel.loc['total_purchase']) * 100
sales_channel.columns = ['Percentage']
print(sales_channel)

                     Percentage
NumDealsPurchases         15.64
NumWebPurchases           27.48
NumCatalogPurchases       17.91
NumStorePurchases         38.96
total_purchase           100.00

# Creating a pie chart to depict the share of different sales channels.
sales_channel = sales_channel.drop(['total_purchase'])
colors = sns.color_palette('pastel')[0:4]
labels = ['Deals', 'Web', 'Catalog', 'Stores']
plt.pie(sales_channel['Percentage'], labels = labels, colors = colors, autopct='%.0f%%')
plt.show()

Q8 The marketing manager wants to understand the performance of different marketing campaigns. Find out which marketing campaign is most successful? Use suitable graphs for visualization.

# Creating a subsection of the dataframe to analyze the marketing campaigns.
campaigns = df.iloc[:, 23:29]

campaigns = campaigns.sum().to_frame()
campaigns.columns = ['Frequency']

campaigns

	Frequency
AcceptedCmp3	163
AcceptedCmp4	167
AcceptedCmp5	163
AcceptedCmp1	144
AcceptedCmp2	30
Response	334

fig = plt.figure(figsize = (10, 5))
labels = campaigns.index.to_list()
plt.bar(labels, campaigns['Frequency'], color ='blue', width = 0.4)

plt.xlabel("Campaigns")
plt.ylabel("Frequency of Accepted Campaigns")
plt.title("Campaigns Accepted")
plt.show()

campaigns2 = pd.DataFrame(df[['AcceptedCmp1', 'AcceptedCmp2', 'AcceptedCmp3', 'AcceptedCmp4', 'AcceptedCmp5', 'Response']].mean()*100, columns=['Percent']).reset_index()

We can also analyze the success rate of the different campaigns them and arrange them accordingly.

px.bar(campaigns2.sort_values('Percent'),x='Percent',y='index',labels ={'index':'Campaign','Percent':'Accepted (%)'},title ='Campaign Acceptance Rate')

Q9 The marketing manager wants to understand which products are performing the best and which are performing the least in terms of revenue. Being an analyst, analyse the data and plot a suitable graph to display a report on revenue generated by different products.

# Creating a subset of the dataset for the different products.
products2 = df[['MntWines', 'MntFruits','MntMeatProducts', 'MntFishProducts','MntSweetProducts' ,'MntGoldProds']]
products2.sum()

MntWines            680816
MntFruits            58917
MntMeatProducts     373968
MntFishProducts      84057
MntSweetProducts     60621
MntGoldProds         98609
dtype: int64

products2.sum().plot(kind='bar',figsize=(6,5),edgecolor=(0,0,0), color=colors, rot=45)

<AxesSubplot:>

average_spend = pd.DataFrame(round(df[[col for col in df.columns if 'Mnt' in col]].mean()), columns = ['Average_Spend']).reset_index()
px.bar(average_spend.sort_values('Average_Spend'),x= 'Average_Spend',y='index',labels ={'index': 'Product',"Average_Spend":'Amount Spent'},title ='Average spending on products')

Q10 The team wants to understand if there’s any pattern between the age of customers and the last campaign acceptance rate. Plot a suitable graph to visualize the distribution of the age with respect to customers who accepted the last campaign.

# Grouping the total acceptance of the last campaign to the ages of the customer.
lastcampaign = df.groupby('age')['Response'].sum()

# Plotting it on a scatter plot
sns.relplot(df['age'],df.groupby('age')['Response'].sum(), data =df)

<seaborn.axisgrid.FacetGrid at 0x255400d0fd0>

lastcampaign = lastcampaign.reset_index()
correlation = lastcampaign['age'].corr(lastcampaign['Response'])
correlation

-0.28636792601520616

Conclusion

• There seems to be little correlation between the age of the customer and acceptance to the last campaign.

Q11 The Chief Marketing specialist wants to visually see which Country has the most number of customers who accepted the last campaign. What is your approach?

# Grouping the dataframe by country and summing the responses.
q11 = df.groupby('Country')['Response'].sum()
q11 = q11.to_frame()
q11

	Response
Country
AUS	23
CA	38
GER	17
IND	13
ME	2
SA	52
SP	176
US	13

# Plotting the above dataset
sns.barplot(x = q11.index, y = q11['Response'])

<AxesSubplot:xlabel='Country', ylabel='Response'>

Conclusion

• We grouped the dataset by country and saw that Spain had the highest number of customers to accepted the last campaign.

Q12 Analyse visually and write your inferences about the relationship between the following.

• Total amount spent Vs Dependents
• Total Purchases Vs Dependents.

# Creating a subsection of the columns we require for this analysis.
dependents = df[['Kidhome', 'Teenhome', 'total', 'total_purchase']]
dependents['Dependents'] = dependents['Kidhome'] + dependents['Teenhome']
dependents = dependents.drop(['Kidhome', 'Teenhome'], axis = 1)

# Comparing the Total amount spent Vs Dependents
d1 = dependents.groupby('Dependents')['total'].sum().to_frame().reset_index()
d1

	Dependents	total
0	0	705647
1	1	533243
2	2	103544
3	3	14554

plt.figure(figsize=(12,4))
sns.barplot(x=d1.Dependents,y=d1.total, palette='Blues')
plt.title('Number of Dependents Vs Total Amount Spent')

Text(0.5, 1.0, 'Number of Dependents Vs Total Amount Spent')

# Comparing the Total Purchases Vs Dependents
d2 = dependents.groupby('Dependents')['total_purchase'].sum().to_frame().reset_index()
d2

	Dependents	total_purchase
0	0	11201
1	1	16334
2	2	5137
3	3	619

plt.figure(figsize=(12,4))
sns.barplot(x=d2.Dependents,y=d2.total_purchase, palette='Blues')
plt.title('Number of Dependents Vs Total Number of Purchases')

Text(0.5, 1.0, 'Number of Dependents Vs Total Number of Purchases')

Q13 Perform Correlation Analysis and write your key inferences

# Selecting the variables for correlation analysis
corr1 = df[['age', 'Income', 'Kidhome', 'Teenhome', 'Recency', 'MntWines', 'MntFruits', 'MntMeatProducts', 'MntFishProducts', 'MntSweetProducts', 'MntGoldProds', 'total', 'NumDealsPurchases', 'NumWebPurchases', 'NumCatalogPurchases', 'NumStorePurchases', 'total_purchase', 'NumWebVisitsMonth', 'AcceptedCmp3', 'AcceptedCmp4', 'AcceptedCmp5', 'AcceptedCmp1', 'AcceptedCmp2', 'Response', 'Complain']]

# Creating a correlation matrix
corr1 = corr1.corr()

sns.heatmap(corr1,cmap='plasma',vmin=-1,vmax=1)
plt.show()

corr2 = corr1.unstack().sort_values()
corr2 = corr2.to_frame().reset_index()

# Checking values of high negative correlation
corr2[corr2[0] < - 0.50] 

	level_0	level_1	0
0	Kidhome	total	-0.56
1	total	Kidhome	-0.56
2	NumWebVisitsMonth	Income	-0.55
3	Income	NumWebVisitsMonth	-0.55
4	NumWebVisitsMonth	MntMeatProducts	-0.54
5	MntMeatProducts	NumWebVisitsMonth	-0.54
6	NumWebVisitsMonth	NumCatalogPurchases	-0.52
7	NumCatalogPurchases	NumWebVisitsMonth	-0.52
8	Kidhome	NumCatalogPurchases	-0.50
9	NumCatalogPurchases	Kidhome	-0.50
10	total	NumWebVisitsMonth	-0.50
11	NumWebVisitsMonth	total	-0.50

# Checking values of high positive correlation
corr2[corr2[0].between (0.60, 0.99)]

	level_0	level_1	0
568	total	MntSweetProducts	0.60
569	MntSweetProducts	total	0.60
570	MntFruits	total	0.61
571	total	MntFruits	0.61
572	MntWines	NumCatalogPurchases	0.64
573	NumCatalogPurchases	MntWines	0.64
574	NumStorePurchases	MntWines	0.64
575	MntWines	NumStorePurchases	0.64
576	MntFishProducts	total	0.64
577	total	MntFishProducts	0.64
578	Income	total	0.66
579	total	Income	0.66
580	total	NumStorePurchases	0.67
581	NumStorePurchases	total	0.67
582	total_purchase	MntWines	0.71
583	MntWines	total_purchase	0.71
584	NumCatalogPurchases	MntMeatProducts	0.72
585	MntMeatProducts	NumCatalogPurchases	0.72
586	total_purchase	NumCatalogPurchases	0.74
587	NumCatalogPurchases	total_purchase	0.74
588	total	total_purchase	0.75
589	total_purchase	total	0.75
590	total_purchase	NumWebPurchases	0.78
591	NumWebPurchases	total_purchase	0.78
592	NumCatalogPurchases	total	0.78
593	total	NumCatalogPurchases	0.78
594	NumStorePurchases	total_purchase	0.82
595	total_purchase	NumStorePurchases	0.82
596	total	MntMeatProducts	0.84
597	MntMeatProducts	total	0.84
598	total	MntWines	0.89
599	MntWines	total	0.89

Inferences

• There seems to be a high negative correlation between the number of kids at home to the amount and number of purchases.
• There seems to be a high negative correlation between income and the number of web visits.
• There is a positive correlation between income and total purchase amount.
• The total purchase amount also has a positive correlation with the number of catalog purchases, which indicates that high value purchases come mostly from the catalog.
• The total number of purchases also has a positive correlation with the number of number of store purchases, which indicates that the most frequency of sales happen at the store.

Q14 Understand the Education background of the customers who complained in the last 2 years. State the Education background of the customers who have registered the most number of complaints.

# Grouping the education levels of the customer according to the total number of complaints.
ed_v_comp = df.groupby('Education')['Complain'].sum().to_frame().reset_index()

ed_v_comp

	Education	Complain
0	2n Cycle	4
1	Basic	0
2	Graduation	14
3	Master	2
4	PhD	1

Conclusion

• The most number of complaints have come from customers who are graduates.

Q15 .Use features 'Total_amount_spent', 'MntFruits', 'MntMeatProducts', 'MntFishProducts', 'MntSweetProducts' and, 'MntGoldProds' in x-axis and y-axis and plot the following plots.

• Plot a pairplot with hue as Response.
• Plot a pairplot with hue as Education.
• Plot a pairplot with hue as Marital Status and write your key observations.

Pairplot with hue as response to the last campaign

sns.set(style="ticks", color_codes=True)
sns.pairplot(df[['MntWines','MntFruits', 'MntMeatProducts', 'MntFishProducts', 'MntSweetProducts', 'MntGoldProds','total', 'Response']], hue = 'Response')

<seaborn.axisgrid.PairGrid at 0x255407e75b0>

Pairplot with hue as Education levels

sns.pairplot(df[['MntWines','MntFruits', 'MntMeatProducts', 'MntFishProducts', 'MntSweetProducts', 'MntGoldProds','total', 'Education']], hue = 'Education')

<seaborn.axisgrid.PairGrid at 0x2554080cf70>

Pairplot with hue as Marital Status

sns.pairplot(df[['MntWines','MntFruits', 'MntMeatProducts', 'MntFishProducts', 'MntSweetProducts', 'MntGoldProds','total', 'Marital_Status']], hue = 'Marital_Status')

<seaborn.axisgrid.PairGrid at 0x255472466a0>